351. Massarsch, K.R. and Fellenius, B.H., Engineering assessment of ground vibrations caused by impact pile driving. Geotechnical Engineering

Transcription

1 351. Massarsch, K.R. and Fellenius, B.H., Engineering assessment of ground vibrations caused by imact ile driving. Geotechnical Engineering Journal of the SEAGS & AGSSEA 46(2)

2 Engineering Assessment of Ground Vibrations Caused by Imact Pile Driving K.R. Massarsch 1 and B.H. Fellenius 2 1 Geo Risk & Vibration Scandinavia AB, Bromma, Sweden 2 Consulting Engineer, Sidney, BC, Canada 1 rainer.massarsch@georisk.se 2 Bengt@Fellenius.net ABSTRACT: Ground vibrations are an imortant design consideration for iles driven by imact hammers. The first task is to determine allowable vibration levels; the second task is to redict the intensity of ground vibrations during driving and the attenuation of ground vibrations with increasing distance. The aer describes the Swedish vibration standard which is alicable for ile driving. Commonly used vibration arameters, associated with the evaluation of vibration measurements, are discussed. The imortance of ile imedance for ground vibrations is highlighted. A simlified calculation method is roosed which can be used to estimate vertical ground vibration velocity as a function of distance from the driven ile. Two case histories have been evaluated and comared with theoretical redictions, using the roosed method of analysis. KEYWORDS: Ground vibrations, Pile driving, Vibration limits, Vibration arameters, Prediction, Frequency, Damage 1. INTRODUCTION Piling is the most widely used foundation method for heavy structures. The geotechnical engineer can choose among a variety of iling methods, but in most cases, driving iles by imact hammer is the most cost-effective alternative. Under unfavourable conditions, driving iles or sheet iles can cause environmental roblems, such as noise, ground movements and vibrations with risk of damage to adjacent buildings or other structures. Therefore, it is imortant for the designer of foundation rojects to be able to redict maximum vibrations, which can be generated due to ile driving. In site of extensive efforts devoted to imrove the understanding of the ile driving rocess, surrisingly few attemts have been made to develo ractical design methods, which can be used to redict ground vibrations as aid in the design and construction rocess. Massarsch and Fellenius (2008) resented a comrehensive concet that encomasses the entire ile driving rocess, addressing the imact of the ile hammer, roagation of stress waves along the ile, and transfer of vibrations from the ile along the shaft and at the toe to the surrounding soil. However, geotechnical engineers need ractical methods to assess ground vibrations, esecially during the reliminary design hase of a roject. To meet this need, a simlified and straight-forward method is resented which does not require extensive theoretical knowledge. In order to evaluate the imact of ground vibrations on the surroundings it is necessary to understand the different arameters which can be used to define ground vibrations. Ambiguity when using definitions of vibration arameters is not uncommon even in the scientific literature. Therefore, definitions of the most imortant vibration arameters are resented. Another imortant task for the design engineer is to comare the redicted level of ground vibrations with vibration limits stated in codes or guidelines. Environmental concerns with resect to noise and ground vibrations can restrict or even rohibit driving of iles. Vibration standards were rimarily develoed for blasting alications, but are also often used to regulate vibrations from construction activities such as ile driving or ground comaction. This aer describes the alication of the Swedish standard which regulates ermissible ground vibrations caused by driving of iles, sheet iles, or ground comaction. An overview of existing international vibration standards alicable to ile driving induced vibrations has been comiled by Massarsch and Fellenius (2014a). 2. DEFINITION OF VIBRATION PARAMETERS The understanding of which arameters can be used to describe vibrations is an imortant requirement when assessing building damage. The following sections describe the most imortant arameters required for evaluating the effect of ground vibrations on buildings and building foundations. A comrehensive discussion of these arameters and their interretation is given by Chameau et al. (1998). 2.1 Vibration Amlitude Vibration amlitude can be defined as the dearture of a oint on a vibrating body from its equilibrium osition. It is equal to one-half the length of the vibration ath. A tyical vibration record from ile driving is shown in Figure 1. The following relationshi exists between different exressions of vibration amlitude. 2 a 2f (2f ) d (1) where a = acceleration, v = article velocity, d = dislacement, and f = vibration frequency. It is thus ossible to derive for sinusoidal vibrations the corresonding amlitude values, when one amlitude value (dislacement, velocity, or acceleration) and the vibration frequency are known. The maximum value of vibration velocity (eak value) occurring during the measuring eriod is in many standards defined as eak article velocity (PPV). If article motions are measured in three orthogonal directions (x, y, and z) simultaneously, it is ossible to calculate the vector sum, v i of the three comonents. v i v v v (2) x1 y1 z1 In the case of sinusoidal vibrations, the average vibration amlitude, x rms (dislacement, velocity, or acceleration) can be exressed by the ratio of root-mean-square (RMS). x rms 1 n x 2 1 x 2 2 x 2 3 x x n where x n etc. are the set of n vibration values. The RMS value, which is frequently used to describe the average vibration intensity, corresonds to the area under the half wavelength. In case of sinusoidal vibrations and only then it is related to the eak amlitude, v eak. v rms 0.7 v eak The relevance of the RMS value deends strongly on the duration of the signal. The so-called CREST factor (eak to (3) (4) 54

3 average) is the ratio between the eak value and the RMS value. In the case of transient vibrations, which are tyically generated by imact ile driving, the duration of the largest motions is small comared to the total length of the signal. For such vibrations, it is ossible to choose the minimum amlitude of interest (i.e. minimum value which is of relevance) and calculate the RMS amlitude from the time that the minimum amlitude is exceeded for the first time to the time when the minimum amlitude is exceeded for the last time in the record. The eak value of the wave is the highest value the wave reaches above a reference, normally zero. This definition is used in the above equations. Note that in engineering alications, frequently the eak-to-trough value (vertical distance between the to and bottom of the amlitude) is used to exress vibration intensity, which ambiguity has caused numerous interretation errors. 2.2 Strain Vibrations assing through material imose strain, which can be calculated from the article velocity and the wave seed. Strain, ε, caused by roagation of a comression wave (P-wave) can be determined from Eq. (5), if the article velocity, v P measured in the direction of wave roagation, and the wave seed, c P are known. (5) c Shear strain, γ can be calculated from the article velocity measured erendicular to the direction of wave roagation, v s and the shear wave seed, c S (Eq. 6). s (6) cs Shear strain is an imortant arameter when assessing settlement in granular soils or disturbance of cohesive soils. Figure 2 Frequency sectrum of the time history shown in Figure 1. The dominant frequency range is indicated A common method of estimating the frequency content (sectrum) of a signal is to erform a Fast Fourier Transformation (FFT). The resulting values are usually resented as amlitude and hase, both lotted versus frequency. A related quantity, which is also widely used to estimate the ower of a signal, is the ower sectrum, which in the frequency domain is the square of FFT s magnitude. Desite its widesread use, there are several limitations associated with the use of Fourier methods for estimating sectra. The Fourier method imlicitly assumes that the signal is stationary. For transient signals such as from imact ile driving and blasting, as well as for many discontinuous signals, this assumtion is not strictly valid. The relationshi between vibration frequency and vibration amlitudes according to Eq. 1 is shown in Figure 3. The black line marks as an examle a vibration velocity of 30 mm/s. At a frequency of 10 Hz, the corresonding dislacement amlitude is 0.48 mm and the acceleration is 0.19 g. If the vibration frequency decreases at constant vibration velocity, the dislacement amlitude will increase. Corresondingly, if the vibration frequency increases at constant vibration velocity, the acceleration amlitude increases. 2.3 Vibration Frequency The time history of the vibration record shown in Figure 1 can be transferred into the frequency domain, Figure 2. The frequency content of a signal is imortant when assessing the effect of vibrations on structures. The simlest method of estimating the dominant frequency is by examining "zero crossings" of the time history. This method works reasonably well for simle, eriodic signals, but is less reliable for comlex, multile-frequency signals. Figure 3 Relationshi for sinusoidal vibrations between frequency and article velocity, acceleration (full lines), and deformation (dotted lines), cf. Eq. (1) 2.4 Wave Length Figure 1 Vertical vibration velocity as function of time. Recording station was on the ground 10 m away from where a recast concrete ile was driven into sandy soil. The ile toe was located 3 m below the ground surface. The value of the eak article velocity (PPV) is indicated (Massarsch and Fellenius 2014) The wave length is an imortant arameter when assessing the risk of damage due to roagation of waves in the ground. The wave length, λ can be determined from the following relationshi. 2 f where: f is the vibration frequency. The largest risk of damage to structures from ground vibrations exists when the wave length corresonds to aroximately the building length, (Massarsch 2000). (7) 55

4 3. DAMAGE POTENTIAL OF VIBRATIONS It is difficult to give general recommendations regarding the damage effects of dislacement, velocity, and acceleration as the sensitivity of structures can deend on many factors. However, the following general observations can be made, which are not generally areciated: Dislacement: at low frequencies, dislacement is the most relevant measure of structural damage as the failure mode is generally due to static stress caused by dislacement (Hooke s law), i.e., damage caused by exceeding material strength. Velocity: indicates how often the dislacement is alied in a given time eriod and is thus related to the fatigue mode of failure (material degradation). As can be seen from Eqs. (5) and (6), strain, which causes material distortion (settlement), deends on vibration velocity and is therefore articularly imortant when assessing settlement in loose granular soils. Acceleration: at high frequencies, the failure mode is normally related to the alied dynamic force caused by inertial forces (Newton s law). It should be ointed out that the above failure modes (stress fatigue dynamic force) can overla and the roer selection of vibration arameter must reflect the tye of roblem. A detailed discussion of damage caused by ground vibrations has been resented by Massarsch and Fellenius (2014b). 4. SWEDISH VIBRATION STANDARD Authorities in many countries are increasingly aware of the imortance of environmental roblems and aly standards more rigorously than in the ast. Vibrations from construction activities are normally not likely to cause damage to buildings or building elements. However, buildings with oor foundation conditions may be very sensitive to vibrations, and, damage may then be instigated or existing cracks and fissures be aggravated. In the case of vibration-sensitive foundation conditions, such as mixed foundations or foundations on loose, granular soils, damage can be caused by vibration-induced increase of re-existing total and/or differential settlements. This asect is not included in most vibration standards. Due to the difficult ground conditions in Sweden, ile driving is used frequently also in vibration-sensitive areas and extensive exerience regarding the effects of driving reformed iles has been accumulated. For these reasons, the Swedish Standard SS , Vibration and shock Guidance levels and measuring of vibrations in buildings originating from iling, sheet iling, excavating, and acking [sic] to estimate ermitted vibration levels was established in The standard which is not widely known outside Scandinavia was articularly develoed to regulate construction activities and is robably the most elaborate standard currently available. It deals with vibrations caused by iling, sheet iling, excavation and soil comaction. Guidance levels of vibrations accetable with resect to otential building damage have been chosen based on more than 30 years of ractical exerience in a wide range of soils. Under the Swedish standard, a risk analysis must be carried out for construction rojects, involving the rediction of maximum ground vibration levels and statement of ermissible vibration levels for different tyes of structures. The roosed vibration values do not take into account sychological effects (noise or discomfort) on occuants of buildings. Neither do they consider the effects of vibrations on sensitive machinery or equiment in buildings. The vibration levels in the standard are based on exerience from measured ground vibrations (vertical comonent of article velocity) and observed damage to buildings, with comarable foundation conditions in Sweden. The vibration level, v, is exressed as the eak value of the vertical vibration velocity. It is measured on bearing elements of the building foundation closest to the vibration source and is determined from the following relationshi. v v 0 F b F m F g (8) where: v 0 = vertical comonent of the uncorrected vibration velocity in mm/s, F b = building factor, F m = material factor and F g = foundation factor. Values for v 0 are given in Table 1 for different ground conditions and construction activities, and are maximum allowable values at the base of the building. It should be noted that in the Swedish standard, the limiting vibration values are indeendent of vibration frequency. The main reason is that within the frequency range of vibrations generated by ile driving and soil comaction, the dominant frequency usually varies within a narrow range (tyically 5 to 30 Hz). Buildings are divided into five classes with resect to their vibration sensitivity cf. Table 2. Classes 1 4 aly to structures in good condition. If they are in a oor state, a lower building factor should be used. Table 1 Uncorrected vibration velocity, v 0 (mm/s) Foundation Condition Clay, silt, sand or gravel Piling, Sheet iling or Excavation Soil Comaction 9 mm/s 6 mm/s Moraine (till) 12 mm/s 9 mm/s Rock 15 mm/s 12 mm/s Table 2 Building Factor, F b Class Tye of Structure Building Factor, F b 1 Heavy structures such as bridges, quay walls, defence structures etc Industrial or office buildings Normal residential buildings Esecially sensitive buildings and buildings with high value or structural elements with wide sans, e.g., churches or museums Historic buildings in a sensitive state as well as certain sensitive historic buildings (ruins) The structural material is divided into four classes with resect to their vibration sensitivity, cf. Table 3. The most sensitive material comonent of the structure determines the class to be alied. Table 3 Material Factor, F m Class Tye of Building Material Material Factor, F m 1 Reinforced concrete, steel or timber 2 Unreinforced concrete, bricks, concrete blocks with voids, lightweight concrete elements Light concrete blocks and laster Limestone, lime-sandstone

5 Table 4 defines a foundation factor. Lower factors are alied to buildings on shallow foundations, whereas buildings on iled foundations are accorded higher factors due to their reduced sensitivity to ground vibrations. Table 4 Foundation Factor, F g Class Tye of Building Material Foundation Factor, F g 1 Slab, raft foundation Buildings founded on friction 0.80 iles 3 Buildings founded on endbearing iles 1.00 The following examle illustrates the ractical alication of the standard: Piles are to be installed in the vicinity of a residential building with brick walls, which are suorted by toe-bearing iles in clay. If the following factors are chosen according to Tables 1 to 4: v 0 = 9 mm/s, F b = 1.00, F m = 1.00, F g = 1.00, the maximum allowable vertical vibration velocity, v, measured at the base of the foundation is 9 mm/s. 5. VIBRATIONS CAUSED BY IMPACT PILE DRIVING When a ile is jacked into the soil, the rocess is usually slow and without vibrations. However, a more effective installation method is ile driving by an imact hammer. The imact of the ile hammer on the ile helmet generates a stress wave that roagates through the ile. Dynamic forces develo along the interface between the ile and the surrounding soil, which can give rise to vibrations. The vibrations roagate in the form of different wave tyes deending on whether the waves are emitted along the ile shaft and/or from the ile toe. Vibrations attenuate with increasing distance from the ile although in some soil layers and buildings, they may actually become amlified due to resonance effects. Vibration roagation caused by ile driving is comlex, as illustrated in Figure 4. The ile hammer (1) transfers the kinetic energy via the ile cushion (2) to the ile (3). The stress wave in the ile generates a dynamic shaft resistance, R M along the ile shaft (4) and dynamic toe resistances, R T at the ile toe (5). Vibrations are generated along the ile shaft and at the ile toe and the distribution of dynamic forces along the ile-soil interface deends on the variation of the dynamic soil resistance. The vibrations roagate in the form of waves to the surrounding soil layers. At the toe, comression waves (P-waves) and shear waves (S-waves) occur, which both extend as sherical waves in all directions. When the waves reach the surface, they are reflected and refracted. The refracted waves are sread as surface waves (R-waves), which roagate with lower attenuation than body waves along the ground surface. A detailed descrition of the dynamic asects of imact ile driving and arising ground vibrations has been resented by Massarsch and Fellenius (2008). In the following, a simlified method of assessing the dynamic ile driving rocess is resented. Figure 4 Proagation of stress wave from imact hammer along the ile and into the surrounding soil and building, (Massarsch 2004). Vibrations are the result of the energy and force imarted by the hammer imact, which both are governed by the hammer imact velocity. The imact velocity can be calculated according to Eq. (8). v 0 2 gh (8) where: v 0 = hammer imact velocity (m/s), g = gravity constant (m/s 2 ), h = hammer height-of-fall (m). It is imortant to realize that the imact velocity which is an imortant arameter for the generation of ground vibrations is indeendent of the hammer mass. The maximum article velocity in the ile that can be generated by the imact can be determined from the hammer and ile imedances according to Eq. (9). v P v 0 1 Z P Z H where: v P = maximum article velocity in ile (m/s), v 0 = hammer imact velocity (m/s), Z P = ile imedance (Ns/m), Z H = hammer imedance (Ns/m). The ile imedance can be determined from Eq. (10). E A Z A c (10) c where: Z P = ile imedance (Ns/m), A P = ile cross sectional area (m 2 ), c P = stress-wave seed in the ile (m/s), ρ = density of the ile material (kg/m 3 ), E P = modulus of the ile (Pa). Equation (9) shows that the article velocity deends on the ratio of the imedances of the ile and the hammer. The hammer imedance is determined similarily. If the hammer and ile imedances are equal, Eq. (9) becomes Eq. (11), indicating that the ile article velocity is half the hammer imact velocity. v P 0.5 v 0 (11) where: v P is maximum article velocity in the ile (m/s), and v 0 is hammer imact velocity (m/s). The dynamic force of the stress wave in the ile, which decides the intensity of the dynamic soil resistance, can be calculated according to Eq. (12). F i Z P v P (12) where: F i = dynamic force in the ile (N), Z P = ile imedance (Ns/m), v P = maximum article velocity (m/s). 5.1 Ground Vibrations due to Pile Driving Ground vibrations are often caused when the ile encounters significant toe resistance. The dynamic toe resistance can be calculated according to Eqs. (13) and (14) (Massarsch and Fellenius 2008). R T 2 v P Z P (13) Z A c s (14) where: R T = toe resistance (N), v P = maximum article velocity (m/s), Z P = imedance of the soil (with resect to P-waves) below the ile toe (Ns/m), A P = ile cross sectional area (m 2 ), c P = = comression wave seed in the soil (m/s), ρ s = the soil bulk density (kg/m 3 ). The comression wave seed in loose, water-saturated soil is about equal to the wave seed in water, but it may be larger in dense or very dense soil. The soil density in most coarse-grained soils ranges from about 1,700 through about 2,200 kg/m 3. Notice that the soil imedance is different to the ile imedance. Combining Eqs. (13) (9) 57

6 Geotechnical Engineeringg Journal of the SEAGS & AGSSEA Vol. 46 No.2 June 2015 ISSN and (14) results in Eq. (15), showing a relation for the dynamic toe resistance as a function of the ile article velocity (the other arameters are nonn variables in a secific case). R T 2 A c s (15) where: R T = toe resistance (N), v P = maximum article velocity (m/s), A P = ile cross sectional area (m 2 ), c P = comression wave seed in the soil (m/s), ρ s = the soil bulk density (kg/m 3 ). 5.2 Emirical Method for Estimating Ground Vibration The engineering ractice has develoed emirical methods for assessing the otential of ground vibrations from ile driving. Equation (16) shows a relation often used. v k W M gh k (16) r r where: v = vertical comonent of the ground vibration (m/s), k = emirical factor (m 2 /s/ Nm), W = imact energy transferred from the hammer to the ile (Nm), r = distance from vibration source to observation oint on the ground surface (m), M = hammer mass (kg), g = gravity constant (m/s 2 ), h = hammer height-of-fall (m). source, which is usually located at some deth below the ground surface. Accordingly, Eq. (16) modifies to Eq. (17). v k Mgh z 2 x 2 (17) where v = vertical comonent of the ground vibration (m/s), k = emirical factor (m 2 /s/ Nm), W = imact energy transferred from the hammer to the ile (Nm), M = hammer mass (kg), g = gravity constant (m/s 2 ), h = hammer height-of-fall (m), z = ile enetration deth, (m), x = horizontal distance from ile to observation oint at ground surface, (m). Figure 6 defines the arameters used in Eq. 17. A common misunderstanding is to consider the distance from the vibration source to the observation oint to be the horizontal distance, x. Instead, the actual distance, r, from the vibration source to the observation oint should be used. Under any circumstance, it is recommended to aly the deth of the vibration source (often the ile toe), r rather than to assume the horizontal distance, x at the ground surface. The emirical factor, k in the equation is not dimensionless, an asect which is imortant when alying the relationshi. Brenner and Viranuvut (1977) reorted results from vibration measurements from ile driving in different soil tyes, Figure 5. Note that Figure 5 shows vibration velocity in mm/s while the distance is given in meters. Often in the literature the average, k = 0.75 is used to estimate ground vibrations. Figure 6 Definition of arameters used in Eqs. (16) and (17) Figure 5 Results of vibration measurements at ile driving with indication of the sread of the k-factor (Brenner and Viranuvut 1977). Note that different units of length are used for article velocity and distance. Figure 5 shows that the sread of the measured values is large, in site of lotting data in a double-logarithmic diagram. Note that the deth of the vibration source (ile enetration deth) is not secified in Eq. ( 16) nor in Figure 5. Another asect which limits the alication of this semi-emirical relationshi is that neither ile characteristics nor the geotechnical conditions are included in the relationshi. When considering vibrations with resect to the distance to the ile, it is imortant to take into account the deth of the vibration Note that Eqs. ( 16) and (17) are based on the energy imarted to the ile. They do not consider loss of energy between the ile hammer and the ile, nor the dynamic soil resistance along the ile shaft and at the ile toe. It can be assumed that the driving energy (hammer mass and height of fall) is well-matched to ensure enetration of the ile. However, if the dynamic soil resistance is smalll and the alied driving energy is larger than required (easy driving), vibration levels calculated according to Eq. (17) will be too high. Thus, Eq. (17) gives an uer limit of ground vibrations generated at the ile toe at mediumm to hard driving conditions. On the other hand, if substantial dynamic soil resistance ("hard" driving) is generated during ile enetration along the ile shaft, cylindrical wave can be generated, which are not considered in Eq. (17). The effect of vibrations emitted along the ile shaft can be estimated according to the rocedure outlined by Massarsch and Fellenius (2008) Heckman and Hagerty (1978) demonstratedd that the k-factor deends on the ile imedance. Massarsch and Fellenius (2008) analyzed their data and found that the k-factor is a linear function of the inverse of the ile imedance, Z P, er Eq. (18), as shown in Figure k P Z (18) where: k = emirical factor (m 2 /s/ Nm), Z P = ile imedance (Ns/m). 58

7 Geotechnical Engineeringg Journal of the SEAGS & AGSSEA Vol. 46 No.2 June 2015 ISSN k-factor (m 2 /s(nm) k = 436 (1/Z P ) INVERSE OF PILE IMPEDANCE, (Z P ) -1 (m/kns) Figure 7 Relationshi between the ile imedance Z P and the emirical coefficient k in Eq. (16), re-analyzed data from Heckman and Hagerty (1978). Note that the constant in Eq. (18) is not dimensionless. Combining Eqs. (16) and (18) results in Eq. (19), which can be used for estimation of ground vibration from ile driving. Figure 9 Variation of vertical ground vibration velocity of hammer with mass of 4,0000 kg, as function of height-of-fall for increasing distance from vibration source (assumed at ile toe) to observation oint at ground surface, according to Eq. (19). In the next examle, the square concrete ile is to be driven to a densee soil layer at 10 m deth. The hammer mass is 4,000 kg and the hammer height-of-fall is assumed to range between 0.2 and 2.0 m. The calculation resultss according to Eq. (19) are shown in Figure 10. Note that in this case, the horizontal distance at the ground surface is shown in linear scale. 436 v P Z F H r W 0 (19) where: v = vertical comonent of the vibration velocity (m/s), k = 436 (m 2 /s/ Nm), Z P = ile imedance (Ns/m), F H = hammer efficacy factor (--), W 0 = nominal energy of ile driving hammer (Nm), and r = distance (m) from vibration source (the ile toe) to an observation oint on the ground surface. 5.3 Samle Calculations The theoretical relations described in the foregoing section can be used in ractice. As an examle, assume a concrete ile with a cross sectional area A P = 0.09 m 2 (0.3 m x 0.3 m), wave seed, c = 3,900 m/s, concrete density, ρ P = 2,630 kg/m 3, which combine to an elastic modulus, E P = 40 GPa. The ile will be driven through a medium dense soil. Eq. (10) gives the ile imedance Z P = 923 kns/m. Figure 8 showss vertical ground vibration velocity vs. distance from source for five hammers used with the same 0.5-m height-of-fall, but having different mass ranging from 3,000 through 7,000 kg. F H = 1.0 was alied to the calculations. Figure 8 Variation of vertical ground vibration velocity as function of hammer mass for increasing distance from vibration source (assumed at ile toe) to observation oint at ground surface, according to Eq. (19). Figure 10 Variation of vertical ground vibration velocity caused by a hammer with mass of 4,000 kg used at a range of height-of-fall, assuming the vibration source (ile toe) is at 10 m deth. 6. CASE HISTORIES Most case histories only rovide information regarding ile tye and vibration velocity as function of horizontal distance from the ile. Only few case histories are available that reort ground vibration velocity with information on ile imed ance and measured ile enetration deth. The measured ground vibration for the following case histories are comared with the semi-emirical vibration measurements during the driving of a series of test iles in southern Sweden, near method resented above. 6.1 Case History - Sweden Nilsson (1989) erformed comrehensive the city of Skövde. A detailed resentation and interretation of the tests was given by Massarsch and Fellenius (2008). The main objective of the vibration measurements was to establish site- would minimize ground vibrations. Ground vibrations were of major secific driving methods and to select the otimal ile tye which concern due to the fact that several buildings and installations in the vicinity were suscetible to vibrations. The soil rofile in the test areaa consisted of about 2 m to 4 m of surface fill, well-comacted, alternating layers of furnace slag sand 59

8 size articles and sand and gravel. Below followed a relatively homogeneous, 12 m thick layer of medium stiff clay with average undrained shear strength of 30 kpa deosited on a layer of sand with a thickness of 7 m on glacial till. Bedrock was encountered at a deth of about 25 m below the ground surface. The groundwater table was located about 3 to 4 m below the ground surface at the to of the clay layer. Unfortunately, data from more detailed geotechnical investigations (such as enetration tests or soil samling) are not available. The existence of a stiff surface layer on to of the clay suggested that vibration roblems would likely occur during the early stage of the driving. Vibration roblems could also be exected during seating of the iles into the bearing layer at 24 to 25 m deth. The ermissible vibration values with resect to damage to the existing structures and installations were estimated according to Swedish standard SS , which has been described above. As the iles were driven into sandy, clayey soils, according to the Swedish standard, a vibration velocity, v 0, equal to 9 mm/s was chosen (Table 1). The buildings were of normal tye (F b = 1), constructed of reinforced concrete (F m = 1.2), and with foundations on toe-bearing iles (F g = 1.0). Therefore, according to Eq. (8), the maximum allowable vibration velocity (the vertical comonent) was v max = 10.8 mm/s. In order to assess whether ile driving would exceed the maximum allowable vertical vibration velocity of 11 mm/s, the driving of two ile tyes reinforced concrete and steel ie iles was monitored by extensive ground vibration measurements (Nilsson 1989) Concrete Piles Three recast concrete iles, called Piles 7, 8, and 10, were driven using an imact hammer. Two of these had the uer ile section covered with an ashalt layer to reduce the otential effect of negative skin friction. Each concrete ile was built u of three segments of lengths ( = 29.3 m), which were connected in the field with a mechanical tye slice. The lower ile segment (starting segment) had a 270 mm x 270 mm square cross section and the uer two ile segments a 235 mm x 235 mm square cross section. The concrete ile had a wave seed of 4,000 m/s, a bulk density of 2,400 kg/m 3 resulting in an elastic modulus of 39 GPa. Thus, the imedances, Z P, of the starting and uer ile elements were 711 and 552 kns/m, resectively.. The iles were driven by a hydraulic hammer tye Banut with a mass of 4,000 kg. During the driving through the overburden soils, the hammer height-of-fall was ket to 0.40 m. It was increased to 0.50 m during the termination driving in the stiff glacial till at a final deth of aroximately 25 m. Figure 11 shows ile enetration resistance (blows/500 mm enetration) as a function of the accumulated (total) number of blows during the driving to 25 m deth to seating the ile in the stiff glacial till. The figure shows the four deths where detailed vibration analyses was carried out. DEPTH (m) Pile Penetration Resistance (Blows/0.5 m) ,200 1,600 2,000 2,400 2,800 3,200 3,600 4,000 Accumulated Number of Blows Figure 11 Pile enetration resistance during driving of the concrete ile with hydraulic hammer to 25 m deth. Also indicated are main soil layers and, by arrows, the deths at which detailed vibration analyses were carried out. (Data from Nilsson 1989). FILL CLAY COARSE- GRAINED SOIL TILL Steel Pie Piles Three steel iles of tye Gustavsberg, made of ductile steel, called Piles 11, 19, and 20, were also installed. Piles 19 and 20 consisted of six segments of 5 m length (total length of 30 m) and Pile 11 had three segments of 5 m length joined by welding in the field. Piles 19 and 20 had diameter 170 mm and wall thickness 13 mm, while Pile 11 had diameter 118 mm and wall thickness 10 mm. Thus, the larger iles had an imedance of 245 kns/m and the smaller ile an imedance of 130 kns/m. The steel ie iles were driven to a total deth of 9.8 m (Pile 19), 19 m (Pile 20) and 25.5 m (Pile 11), using an imact hammer with mass of 1,500 kg and 300-mm height-of-fall. During a brief testing hase, the 4,000 kg Banut hammer was also used at the height-of-fall ranging from 100 to 300 mm. It was observed that during the easy driving with the heavy ile hammer, ground vibrations were lower, as oosed to when it was driven with the lighter hammer Vibration Measurements Concrete Pile. Vibration measurements were erformed using five geohones laced at 10, 20, and 40 m distance from the resective test ile. A data logger recorded the eak value of vibration velocity at each hammer blow as well as the deth of the ile at each measurement. Figure 12 shows the vertical vibration velocities measured at 10 (V1), 20 (V2), and 40 (V3) m horizontal distances from Pile 10 as function of selected ile toe deths. DEPTH TO PILE TOE (m) VERTICAL VIBRATION VELOCITY (mm/s) V3 V2 V1 h = 0.4m h = 0.5m Figure 12 Vertical vibration velocities measured when driving concrete Pile 10. The measurements were taken at 10 (V1), 20 (V2), and 40 (V3) m horizontal distance as function of ile enetration deth. The hammer height-of-fall is indicated as h. (Data from Nilsson 1989). When the ile was driven through the surface fill, the magnitude of the vibration amlitudes at 10 and 20 m distance are relatively equal, comared to that at 40 m distance. The vertical vibration velocity decreases markedly with increasing horizontal distance from the ile. At a ile deth range of 17 to 25 m, the direct distances from the ile toe to the measurement oints V1 and V2 were 26 m and 32 m, resectively. The distance difference is small in terms of vibration roagation, which exlains why the measured vibration amlitudes are almost the same. Figure 13 shows the results of vibration measurements for the three comosite concrete iles, driven with imact hammer (4,000 kg) and 0.4 m height-of-fall (raised to 0.5 m during the termination driving). The vibration velocity curves (one dashed and one dotted), calculated by means of Eq. (19) for the two ile imedances, are shown as well as the measured values of vertical vibration velocity. F H = 1.0 was alied to the calculations. 60

9 That is, the driving energy was larger than needed, that is, the ratio of actual energy in the wave down the ile was smaller than usual. The dashed lines in the figure show vibration velocities calculated for a lower driving energy (hammer mass 400 kg) with reduced height-of-fall (0.1 and 0.3 m) used during easy driving. The calculated values for the low energy driving lie closer to the measured values. Another exlanation of the discreancy between calculated and measured vibration amlitudes could be that the relationshi shown in Figure 7 does not aly correctly to iles with very low imedance. Note that in Figure 7, only one data oint was in the range of low ile imedance. Additional data for iles with low imedance are needed to substantiate the validity of Eq. (19) for iles with very low imedance. Figure 13 Vertical vibration velocities measured when driving the comosite concrete iles lotted together with two values of ile imedance calculated from Eq. (19). (Data from Nilsson 1989). The calculated lines lie above the measured values with the excetion of measuring oint V3 at 40 m distance, where vibrations during driving of Pile 10 exceeded the calculated values in a few instances. However, the Eq. (19) calculations agree surrisingly well with the measured ground vibrations, considering the comlexity of the roblem. Note that the diagrams in Figure 13 is in linear scale, which gives a more realistic comarison between measured and calculated values as oosed to a diagram showing values in log-log scale, otherwise commonly used for reorting vibration measurements Vibration Measurements Steel Pie Pile The results of vibration measurements during driving of the steel ie iles using a Banut hydraulic hammer are shown in Figure 14. The three steel ie iles were generally driven with a 1,500 kg hammer and 0.3 m height-of-fall. F H = 1.0 was alied to the calculations. 6.2 Case History - Thailand Brenner and Chittikuladilok (1975) measured vibrations due to driving recast concrete iles in Bangkok clay at two sites called Lak Si, located north of Bangkok, and EGAT, located south of Bangkok. The iles were recast concrete iles driven to deths ranging between 18 and 28 m. A large number of measurements were carried out at the ground surface at different distances from the ile during ile enetration. The aer also includes information on ile cross-section and ile material which is required information for the calculation of ground vibrations according to the above concet Site Conditions At both sites, the ground surface was raised by an aroximately 2 m thick fill consisting of sand (Lak Si) and sand and gravel (EGAT). Below the fill, the rofile consisted of very soft to soft Bangkok clay followed at a deth of 14 to 15 m by stiff Bangkok clay. At the EGAT site, an aroximately 3 m thick layer of loose, fine clayey sand was found, interbedded in the stiff clay between deths about 6.5 through 9.5 m deths. Cone enetration tests were erformed at both sites and Figure 15 shows the distribution of cone stress, q c. Note the high cone enetration resistance at the EGAT site at about 8 m deth. Figure 14 Vertical vibration velocity measured when driving the steel ie iles lotted together with values calculated from Eq. (19). Also shown are vibration velocities calculated for driving with reduced energy: hammer mass of 400 kg and height-of-fall 0.1 m. (Data from Nilsson 1989). It is aarent that in the case of steel ie iles, the vibration velocities calculated by Eq. (19) are larger than the measured ground vibrations and esecially so for the ile with smaller diameter (Pile 11). In the oinion of the authors, the main reason for the measured vibration velocities being smaller than those calculated, is the easy enetration of the small steel ie iles when being driven by a heavy hydraulic hammer. Figure 15 Distribution of cone stress, qc, at two Bangkok test sites. (Data from Brenner and Chittikuladilok 1975). At Lak Si, concrete iles with a ile area of m 2 were installed. The density of the concrete iles was assumed to 2,440 kg/m 3. The elastic modulus of the concrete iles was 39 GPa. The wave seed was 4,000 m/s. The ile imedance of the concrete iles was thus 658 kns/m. At the EGAT site, concrete iles with a m 2 ile area were installed. The density of the concrete iles was assumed to 2,440 kg/m 3. A 39 GPa elastic modulus and a 4,000 m/s wave seed were assumed for the concrete iles. Table 5 rovides a summary of the ile and hammer data for the two sites. 61

10 Table 5 Summary of Pile Driving Information Site Lak Si EGAT Pile cross section (m 2 ) Pile length (m) Pile imedance (kns/m) 658 1,755 Mass of hammer (kg) 4,700 7,000 Hammer height of fall (m) the ile shaft at aroximately. In Eq. (18), the source of vibrations has been assumed to be located at the ile toe. Figure 17 shows the measured vertical vibration velocity at lateral distances ranging from 1 to 20 m from the ile. Driving the concrete ile through the hard sand fill layer close to the ground surface would have risked breaking the iles. Therefore, a steel robe was used to re-bore through the fill. No information is available about the steel robe and no vibration measurements were reorted from the driving through the fill layer. Desite the reboring, the concrete iles had to be driving through the sand fill. During ile driving, vertical vibration velocity was measured at the ground surface at several distances from the resective ile. The ile enetration deth was recorded for each measurement record. The following general observations were reorted by Brenner and Chittikuladilok (1975): A sudden decrease in vibration occurred when the ile toe enetrated from the surface sand fill into the soft clay layer. At the EGAT site, a significant increase in vibration took lace when the ile toe encountered the layer of fine sand at deth between 6.5 and 9.5 m, roducing there a relative maximum in ground vibration velocity. When the ile toe reached the stiff clay layer, a ronounced increase in vibration level occurred due to the greater enetration resistance. A distinct increase in vibration intensity from driving through the stiff clay layer, however, aeared to occur only at surface oints away from the ile. The results of ground vibration measurements and calculated vibration levels according to Eq. (19) are shown in Figure 16 for the Lak Si site. Vertical ground vibrations were measured at four distances at the ground surface (10, 18, 25, and 30 m). Figure 17 EGAT Site: vertical vibration velocity measured during driving of concrete iles, driven with imact hammer (mass 7,000 kg) and 1.0-m height-of-fall. Also shown are the calculated vibration velocities according to Eq. 18. (Data from Brenner and Chittikuladilok 1975). At the EGAT site, the general trend of measured ground vibrations as function of distance to the ile toe is good. Close to the ile, measured ground vibrations are in very good agreement with calculated values according to Eq. (18). At increasing distance, measured values are somewhat higher than those redicted. This effect can be exlained by the contribution of vibrations generated along the ile shaft at about 7 m deth, adding to the vibrations generated at the ile toe. Also, surface waves can become more imortant at larger distance from the vibration source, an effect which is not included in the analysis according to Eq. (18). However, in general, measured vibration attenuation is in good agreement with the roosed method of analysis. Figure 16 Lak Si Site: vertical vibration velocity measured during driving of concrete iles, driven with imact hammer (mass 4,700 kg) and height-of-fall of 0.5 m. Also shown are the calculated vibration velocities according to Eq. (19). (Data from Brenner and Chittikuladilok 1975). The measured vibration velocities are in general agreement with those calculated according to Eq. (18), although at shorter distance, measured ground vibrations are slightly higher than those calculated. One likely reason for this difference between measured and calculated values could be that a significant art of the vibration energy was generated by the dynamic soil resistance acting along 7. SUMMARY AND CONCLUSIONS When lanning and designing a iling roject, the geotechnical engineer must answer two questions: how strong will the maximum ground vibrations be during ile driving and which vibration levels are accetable for the secific case. The definition of vibration arameters is an imortant asect when evaluating vibration measurements. Therefore, relevant arameters are described and defined. A method of defining ermissible ground vibrations from imact ile driving according to the Swedish vibration standard is resented. The standard takes into consideration the ground conditions, building standard and foundation conditions. The standard is used in the Nordic countries of Euroe with good results. It should be noted that the geological conditions and housing standards in Sweden which can have imortance for the dynamic resonse of buildings may not be alicable in countries with significantly different foundation conditions and construction methods. This asect needs to be taken into consideration when alying the Swedish vibration standard. The concet is resented which describes the factors, which are of significance for the generation of ground vibrations due to imact ile driving. A simle aroach is roosed for the rediction of vertical ground vibrations, which takes into account ile driving energy (hammer mass and height-of-fall) and ile imedance. Based on limited data, an almost linear relationshi was found between the inverse of ile imedance and the k-factor. Figure 7 shows that the 62

11 k-factor deends on the inverse of the ile imedance. This asect of analysing ground vibrations due to imact ile driving has reviously not been considered in a quantitative way. Prediction of ground vibrations caused by imact ile driving is a comlex task. The objective of the aer is not to redict the ground vibrations during all hases of ile driving. Rather, emhasis has been laced on redicting uer limits of vertical ground vibrations in the near-field, i.e., to a distance equal to about two ile lengths. A method is resented which makes it ossible to estimate the vertical comonent of ground vibrations during hard driving, taking into consideration the imortance of ile imedance. However, the roosed method is based on simlified assumtions and needs to be verified by field measurements. Equation (19) defines the arameters on which the rediction of vertical ground vibrations in the near-field is based. Vibration attenuation is redicted, assuming that the distance from the source of dynamic driving resistance below the ground surface to a oint at the ground surface is known. Thus, the actual distance from the energy source to a oint on the ground surface should be used as the distance (r), and not the horizontal distance at the ground surface. It should be noted that vibrations can be emitted at the ile toe as well as along the ile shaft. Geotechnical investigations are needed to determine where along the ile vibrations are rimarily emitted. In the resent method, it is assumed that the rimary source of vibration is at the ile toe. Samle calculations have been resented which illustrate the effect of hammer mass and hammer height-of-fall on ground vibrations as function of distance from the vibration source. Two case histories with very different ground conditions, ile tyes, and driving methods were analyzed. The test objective was to determine an uer boundary of ground vibration velocity, which can occur during the driving rocess. The agreement between calculated and measured vertical ground vibrations is reasonable, considering the comlexity of the roblem. Vibration attenuation has been shown in linear scale, which give a better understanding of the accuracy of vibration attenuation than logarithmic diagrams. The results from the evaluation of case histories confirm that the general trend of vibration attenuation is catured by the simle relationshi given by Eq. (18). The ile imedance is an imortant arameter. The main conclusion of the ile driving tests is that Eq. (18) surrisingly well redicts ground vibrations generated by driving concrete iles, and moderately well in the case of steel ie iles with significantly lower imedance. The values calculated according to Eq. (18) are generally considered to give conservative results. It is imortant to oint out that the method does not consider the variation of dynamic soil resistance along the ile toe and ile shaft, which is ossible when alying the more comlex concets roosed by Massarsch and Fellenius (2008). Therefore, when secial care is needed, the assessments should be verified and adjusted by results of field measurements. Esecially in the case of easy ile driving (low ile enetration resistance), the roosed method will overestimate actual ground vibrations. 8. ACKNOWLEDGEMENT The authors wish to acknowledge the valuable comments and criticism by the two reviewers. Also, the encouragement by Prof. Balasubramaniam to reare the aer is gratefully acknowledged. 9. REFERENCES Brenner, R.P. and Chittikuladolik, B., Vibrations from ile driving in the Bangkok area. Geotechnical Engineering, 6(2), Brenner, R.P. and Viranuvut, S., Measurement and rediction of vibration generated by the dro hammer iling in Bangkok subsoils. Proceedings of the 5th Southeast Asian Conference on Soil Engineering, Bangkok, July 1977, Chameau, J-L., Rix, G.J. and Emie, L., Measurement and analysis of civil engineering vibrations. Fourth International Conference on Case Histories in Geotechnical Engineering. St. Louis, Missouri, March 8-15, Heckman, W.S. and Hagerty, D.J., Vibrations associated with ile driving. American Society of Civil Engineering, Journal of the Construction Division, 104(CO4) Massarsch, K.R., Settlements and damage caused by construction-induced vibrations. Proceedings, Intern. Worksho Wave 2000, Bochum, Germany December 2000, Massarsch, K.R., Vibrations caused by ile driving. Dee Foundations Institute Magazine. Part 1: Summer Edition, , Part 2: Fall Edition, Massarsch, K.R. and Fellenius, B.H., Ground vibrations induced by ile driving. 6th International Conference on Case Histories in Geotechnical Engineering, Edited by S. Prakash, Missouri University of Science and Technology, August 12-16, 2008, Arlington, Virginia, Arlington, VA, August 11 16, Keynote lecture. 38. Massarsch, K.R. and Fellenius, B.H., 2014a. Ground Vibrations from Pile and Sheet Pile Driving Review of Vibration Standards. Proceedings, DFI/EFFC International Conference on Piling and Dee Foundations, Stockholm, May 21 23, 2014, 15. Massarsch, K.R. and Fellenius, B.H., 2014b. Ground Vibrations from Pile and Sheet Pile Driving Building Damage. Proceedings, DFI/EFFC International Conference on Piling and Dee Foundations, Stockholm, May 21 23, 2014, 15. Nilsson, G., Markvibrationer vid ålslagning (Ground vibrations during ile driving). Examensarbete Nr. 3:89. Det. of Soil and Rock Mechanics, Royal Institute of Technology (KTH). Stockholm, Sweden, 43. and Aendix. - Swedish Standard., Vibration and shock Guidance levels and measuring of vibrations in buildings originating from iling, sheet iling, excavating and acking to estimate ermitted vibration levels. SS Swedish Institute for Standards, SIS. Stockholm 1999, 7. 63